Higher Order Aberrations


I'm a bit of scientific jack of all trades. I'm a geek. I like cats.
Consider yourself warned.
mityelpoc: leelacson:photojojo: Mistakes and glitches can make some awesome photographs, and can really make us think about how photography works. This crazy flying rainbow plane on google maps (click Satellite option to view it) is just one example we are in awe of. Anyone have any idea how or why this happened? We’d love to know! (via sarahpalmer, lane, & @lamartin) OK… so, this happens because of the way Google’s satellite photos are taken (I think). The satellite that takes a lot of Google Earth’s Satellite based (high resolution) images is called QuickBird. QuickBird uses a linear array of CCDs. One set takes panchromatic images (grey scale) and the second set takes multi spectral images (MSI). The MSI imager is mounted in different physical locations on the satellite, so its images are acquired at different times. These images (grey and MSIs) are time re-aligned on the ground on the assumption that there are no moving objects in the field of view. If there are, you get something like what you see here. You’re right about the reason this happens, but your details are just a little off.  There are four CCDs, one for panchromatic (grey scale) and three others with individual color filters over them that generate the red, green, and blue.  (There’s also a fifth CCD in the NIR, but Google Maps doesn’t care about that.) They use what’s called a push-broom scanning process, and the MSI sensors are usually scanning behind the pan sensor (in time). Why do it this way?  Because the pan sensor has more light available (no filters), it can have smaller pixels to generate the same signal.  Smaller pixels translates to higher resolution on the ground.  This high-res image is then combined with the lower-res color image to produce the final image you see in Google Maps.  The resolution difference is about a factor of 3, according to the Quickbird specs. The CCDs are also not linear arrays but 2D arrays, although they don’t produce a 2D image.  Instead, the transfer of charge in the CCD is synced to the speed of motion of the image across the CCD.  This is called time delay and integration, and gives better single-to-noise ratio in the images (the SNR goes up as the square-root of the number of TDI stages; i.e. 16 TDI gives you 4X improvement in the SNR). WorldView-2 actually has eight multispectral bands.

mityelpoc:

leelacson:photojojo:

Mistakes and glitches can make some awesome photographs, and can really make us think about how photography works.

This crazy flying rainbow plane on google maps (click Satellite option to view it) is just one example we are in awe of. Anyone have any idea how or why this happened? We’d love to know!

(via sarahpalmer, lane, & @lamartin)

OK… so, this happens because of the way Google’s satellite photos are taken (I think). The satellite that takes a lot of Google Earth’s Satellite based (high resolution) images is called QuickBird. QuickBird uses a linear array of CCDs. One set takes panchromatic images (grey scale) and the second set takes multi spectral images (MSI). The MSI imager is mounted in different physical locations on the satellite, so its images are acquired at different times.

These images (grey and MSIs) are time re-aligned on the ground on the assumption that there are no moving objects in the field of view. If there are, you get something like what you see here.

You’re right about the reason this happens, but your details are just a little off.  There are four CCDs, one for panchromatic (grey scale) and three others with individual color filters over them that generate the red, green, and blue.  (There’s also a fifth CCD in the NIR, but Google Maps doesn’t care about that.) They use what’s called a push-broom scanning process, and the MSI sensors are usually scanning behind the pan sensor (in time).

Why do it this way?  Because the pan sensor has more light available (no filters), it can have smaller pixels to generate the same signal.  Smaller pixels translates to higher resolution on the ground.  This high-res image is then combined with the lower-res color image to produce the final image you see in Google Maps.  The resolution difference is about a factor of 3, according to the Quickbird specs.

The CCDs are also not linear arrays but 2D arrays, although they don’t produce a 2D image.  Instead, the transfer of charge in the CCD is synced to the speed of motion of the image across the CCD.  This is called time delay and integration, and gives better single-to-noise ratio in the images (the SNR goes up as the square-root of the number of TDI stages; i.e. 16 TDI gives you 4X improvement in the SNR).

WorldView-2 actually has eight multispectral bands.

  1. deepseaian reblogged this from lane
  2. yoursplusmine reblogged this from pifkie
  3. pifkie reblogged this from photojojo
  4. monochromecyclone reblogged this from photojojo
  5. britishnigga reblogged this from ekseksekso
  6. ekseksekso reblogged this from photojojo
  7. apathylife reblogged this from photojojo
  8. intosomethingbeautiful reblogged this from photojojo
  9. -tea reblogged this from micahbaldwin
  10. jackbarapuppy reblogged this from imxgen
  11. imxgen reblogged this from photojojo and added:
    it happened, it looks pretty (:
  12. ophelia-looking-glass reblogged this from photojojo
  13. megandecorp reblogged this from photojojo
  14. sourskin reblogged this from photojojo
  15. sleeping-asianprincess reblogged this from photojojo
  16. thejesusofsuburbia reblogged this from photojojo
  17. mako-28 reblogged this from pdl2h
  18. thegaytrifecta reblogged this from photojojo
  19. grinandshare reblogged this from photojojo
  20. iamdineshdharma reblogged this from photojojo
  21. ultrafastx reblogged this from mityelpoc and added:
    You’re right about the reason this happens, but your details are just a little off. There are four CCDs, one for...
  22. lexburg reblogged this from photojojo